Application of texture analysis of CT and MR images to determine the histologic grade of hepatocellular cancer and it’s differential diagnosis: a review

Abstract

In recent years, more foreign publications are devoted to the use of texture analysis or radiomics in solving certain diagnostic problems, including the diagnosis of hepatocellular cancer (HCC). This method of processing medical images allows for a comprehensive assessment of the structure of neoplasms by extracting a large number of quantitative features from medical images.The purpose of the study was to determine the role of texture analysis of CT and MR images in differential diagnosis and determination of the degree of differentiation of HCC based on a review and analysis of the results of publications.We searched for scientific publications in the PubMed information and analytical system for 2015–2021. by keywords: “HCC”, “texture analysis” (texture analysis), “radiomics”, “CT”, “MRI”, “grade”, “differential diagnosis”. After excluding reviews of publications and studying the full text of articles, 21 articles were selected for analysis.Despite the growing number of publications devoted to the successful use of textural analysis of CT and MR images, including non-invasive assessment of the histological grade of HCC and in the differential diagnosis of HCC with hypervascular neoplasms, metastases, regenerative and dysplastic nodes, the use of such type of analysis in routine practice is limited due to the lack of standardized methods for performing texture analysis, which leads to low reproducibility of the results. The parameters of image acquisition and methods of image preprocessing and segmentation affect the reproducibility of the obtained texture features. In addition, the presented studies were performed using different MR sequences and phases of contrast enhancement, as well as different software, which makes it difficult to compare the obtained data.The use of texture analysis certainly demonstrates promising results and requires further investigation to systematize and standardize the obtained data in order to develop an optimal diagnostic model for wide clinical use.